Modified DESeq2 plotPCA function with sample names and proportion of variance added. Sample names will be shown underneath each dot. The axis will display proportion of variance for each principal component. Tested using DESeq2 1.2.8, 1.6.2, and 1.8.1. The DESeq2 plotPCA function switched from lattice to ggplot2 in version 1.5.11.

plotPCAWithSampleNames.R

plotPCAWithSampleNames = function(x, intgroup="condition", ntop=500)
{
	library(RColorBrewer)
	library(genefilter)
	library(lattice)

	# pca
	rv = rowVars(assay(x))
	select = order(rv, decreasing=TRUE)[seq_len(min(ntop, length(rv)))]
	pca = prcomp(t(assay(x)[select,]))

	# proportion of variance
	variance = pca$sdev^2 / sum(pca$sdev^2)
	variance = round(variance, 3) * 100
	
	# sample names
	names = colnames(x)

	# factor of groups
	fac = factor(apply(as.data.frame(colData(x)[, intgroup, drop=FALSE]), 1, paste, collapse=" : "))

	# colors
	if( nlevels(fac) >= 10 )
		colors = rainbow(nlevels(fac))
	else if( nlevels(fac) >= 3 )
		colors = brewer.pal(nlevels(fac), "Set1")
	else
		colors = c( "dodgerblue3", "firebrick3" )

	# plot
	xyplot(
		PC2 ~ PC1, groups=fac, data=as.data.frame(pca$x), pch=16, cex=1.5,
		aspect = "fill",
		col = colors,
		xlab = list(paste("PC1 (", variance[1], "%)", sep=""), cex=0.8),
		ylab = list(paste("PC2 (", variance[2], "%)", sep=""), cex=0.8),
		panel = function(x, y, ...) {
			panel.xyplot(x, y, ...);
			ltext(x=x, y=y, labels=names, pos=1, offset=0.8, cex=0.7)
		},
		main = draw.key(
			key = list(
				rect = list(col = colors),
				text = list(levels(fac)),
				rep = FALSE
			)
		)
	)
}